Optimal Stealthy Attacks on Industrial Cyber Physical Systems Under Random Access Protocol

Abstract

This article designs the stealthy attack against industrial cyber-physical systems under the random access protocol (RAP), which is used to avoid data collisions. The aim is to design an optimal attack that degrades the estimation performance while remaining strictly stealthy. Under the RAP, only a part of the transmitted data can be utilized to construct attack signals at each time instant, which brings challenges to attack design. To obtain the precondition for attack design, a novel modified algebraic Riccati equation (MARE) is derived, where the MARE’s convergence problem is analyzed and a sufficient condition for the existence of the modified Kalman filter’s steady state is presented. Then, the optimal attack strategy is obtained by solving a convex optimization problem. Furthermore, the analytical optimal attack parameters are presented and proved. Finally, experiments and simulations are performed to verify our results.